Stress, memory, and the predictive brain – 2026 Dirk Hellhammer Award

Stress-induced changes in learning and memory facilitate adaptation to current stressors and to similar future challenges. A key element of these adaptive changes is the enhanced memory for stressful episodes themselves, which can extend to events encoded shortly before or after the stressor. Drawing on the predictive brain framework, I argue here that these memory enhancements primarily serve to update models of the environment. Accordingly, memory should not be globally and uniformly enhanced for all stimuli encoded under stress and in close temporal proximity. Instead, there should be more selective mechanisms of memory formation, guided by the unexpectedness or prediction error (PE) associated with events, as high PE events are particularly relevant for model updating. Emerging evidence supports this perspective, indicating that the memory boost for stressful episodes and nearby events is at least partly driven by the evoked PE. This PE-centred view highlights the adaptive cognitive mechanisms underlying memory formation under stress and may have important implications for modifying memory distortions in stress-related mental disorders.